A Characterization of the Dirichlet Distribution with Application to Learning Bayesian Networks
Artificial Intelligence
2013-02-21 v1 Machine Learning
Abstract
We provide a new characterization of the Dirichlet distribution. This characterization implies that under assumptions made by several previous authors for learning belief networks, a Dirichlet prior on the parameters is inevitable.
Cite
@article{arxiv.1302.4949,
title = {A Characterization of the Dirichlet Distribution with Application to Learning Bayesian Networks},
author = {Dan Geiger and David Heckerman},
journal= {arXiv preprint arXiv:1302.4949},
year = {2013}
}
Comments
Appears in Proceedings of the Eleventh Conference on Uncertainty in Artificial Intelligence (UAI1995)